I recently posted a significant update of Scriblio to SVN trunk. The  
code is alpha quality and not ready for production, but it includes a  
number of enhancements, some incremental, some more significant.

More immediately, however, this marks a change in how the SVN  
repository is organized. The code in /trunk is now in some stage of  
development, stable releases can be found in /tags.

Development:
http://svn.scriblio.net/plugin/trunk

Stable:
http://svn.scriblio.net/plugin/tags/2.3v6
http://about.scriblio.net/downloads/scriblio_plugin.tar.gz

So, what's in the development version?

The main database query code has been rewritten for better performance  
(though support for Sphinx has not yet been significantly tested)

The taxonomy registration is now more efficient, and makes way for us  
to get pretty permalinks for queries like site.org/browse/subject/
history rather than /browse/?subj=history as it is now. The logic for  
creating links hasn't been updated to create links in that form, and  
the admin panel doesn't yet let you configure it, but did I say this  
was alpha-quality?

It supports/requires bSuite 4 (bSuite 4 is another project of mine,  
and it's where I implement code that can be used more broadly). bSuite  
4 is a significant rewrite on its own, and brings with it some  
significant new features: http://maisonbisson.com/blog/post/12142/bsuite-4-public-beta

One of the features that Scriblio uses in bSuite is the "related  
items" functionality. And one of the most exciting _potential_ new  
features going forward may be offering related titles based on both  
metadata (as it currently does) and user activity. bSuite tracks  
various items viewed in a session (sessions expire after 30 minutes of  
inactivity) and in my testing, the relationships shown from the data  
are very interesting.

A very significant change for those running current versions is how  
links in a post's content are generated. The new version takes  
advantage of WordPress 2.5+'s new "shortcode" feature. This adds  
additional functionality and makes Scriblio content more portable, but  
it also means that current users will need to re-harvest their  
collections to upgrade.

But current users will want to re-harvest their collections anyway, at  
least III users will, to take advantage of K.T. Lam's (of HKUST)  
improved III harvester. The resulting records are cleaner, have more  
metadata, and offer more facets. Thank you K. T. Lam and HKUST!

Also related to record harvesting, if a bookjacket is not available  
for a record, it uses an API to generate one using the title and  
author information from the record. These "fakejackets" are stored on  
your Scriblio server and require that the wp-content/uploads folder be  
writable, but it also means that we'll no longer suffer the  
placeholder bookjacets.

I think the table creation problem is fixed. I think.

In the land of small changes, you'll notice that RSS feeds for catalog  
content now have the correct author information set, rather than  
simply "cataloger."

The list of things that don't yet work include the MARC harvester (and  
all the other harvesters besides the III one) and probably a few other  
things.

I don't recommend this code for most people, but if you want to  
explore, take a look at it. Tell us about what you find on the list.

--Casey

On 8 Haziran, 03:53, Casey Bisson <ca...@scriblio.net> wrote: